A Generative AI-based climate-resilient agriculture framework is proposed to enhance crop yield forecasting and encourage adaptive farming. The goal is to develop a strong and data-driven system that enables sustainable crop planning in various Indian states. The scope includes the forecasting of yield rates for a broad array of crops over a 50-year period. Initially, an XGBoost model trained on actual datasets of crop yield, temperature, rainfall, and soil nutrients produced low accuracy. To overcome this shortcoming, Synthetic data were generated using Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) to fill data gaps and enhance prediction, which symbolized future agro-climatic conditions. The augmented dataset tremendously boosted model performance. A T5 language model also enriches the system with bilingual (English and Tamil) farming advice from input conditions. The system also possesses a user-friendly interface where the users input crop/environmental parameters and get the predicted yield and recommendations for optimal cultivation. Validation revealed improvement, where the ensemble model gained accuracy of 90%, precision of 88%, recall of 89% and an F1 score of 88.5%. This reflects the potential of integrating generative models and predictive analytics in agriculture. Further research may include, integration of real-time satellite and weather data, inclusion of more crop variables and the development of mobile applications enabling widespread real-time utilization by farmers.
F et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: